Mobile-native local AI Agent framework for phones and edge devices, with an iOS runtime, fully offline local path, on-device models, native iOS Skills, and optional Mac Gateway inference

OpenClaw alternative in your pocket

A tool for converting ONNX files to LiteRT/TFLite/TensorFlow, PyTorch native code (nn.Module), TorchScript (.pt), state_dict (.pt), Exported Program (.pt2), and Dynamo ONNX. It also supports direct conversion from LiteRT to PyTorch.

PokeClaw (PocketClaw) — first on-device AI that controls your Android phone. Gemma 4, no cloud, no API key. Poke is short for Pocket.

The most advanced, fully offline client-side AI suite on Android today.

Flutter plugin for Ultralytics YOLO

XR Blocks is a lightweight WebXR + AI library for rapidly prototyping advanced AI + XR experiences.

Run local LLMs like Gemma, Qwen, and LLaMA on Android for offline, private, real-time chat and question answering with LiteRT and ONNX Runtime.

Turn your Android phone into an OpenAI-compatible LLM inference server — Fully local, private and Open Source

Run ml models with both tflite & litert runtimes on edge devices

On-device VAD / streaming STT / TTS / diarization in C++17 (ONNX + LiteRT) with a voice-agent pipeline. Linux, Windows, Android.

Swift package for running LiteRT-LM models on iOS. Wraps Google's C API in a clean, async/await Swift interface.

This repository is an implementation of converting the YOLOv10 object detection model to LiteRT (.tflite) format and deploy it on Android using Google AI Edge for on-device inference.

A desktop application that estimates where the user is looking on the selected display and renders a red gaze marker at the estimated hit position. MediaPipe dependency-free.

on-device sensitive content blocker for Android. Works across any app powered by a custom-trained on-device AI model.

Pocket AutoML: Tutorial for Creating an Android App for Image Classification with Deep Learning

Image Classification, Image Captioning and LLM inference with LiteRT

A clean, private calculator with an optional on-device AI math solver — no tracking, no ads.

A set of tools and examples for converting and utilizing powerful vision models, DINOv3 and EdgeTAM (SAM2), within the ONNX ecosystem.

Graph-based C++ runtime for building and executing AI, ML, and computer vision pipelines across devices.